Illustration of number of dummy variables required to model parameters of effectiveness
Supplemental material to “Everything should be as simple as possible, but no simpler: towards a protocol for accumulating evidence regarding the active content of health behaviour change interventions” by Peters, de Bruin & Crutzen in Health Psychology Review (2014)1 of 2
Tables 1 and 2 illustrate the possible combinations of the conditions for effectiveness of fear appeals
and modelling. For fear appeals, of the 16 combinations, the first four (1-4) do not contain any
manipulation of threat, and so are unlikely to be considered applications of fear appeals. This leaves 12
types of the fear appeal method. Similarly, for modelling, of the 16 combinations, let us ignore the first
8 combinations that are not attended to or remembered as this is not under control of the intervention
developer. Thus, to assess the effectiveness of fear appeals and modelling using a meta-regression
model, we will need 4 predictors to represent 12 types of fear appeals and 3 predictors to represent 8
types of modelling. However, a regression model with only those seven predictors would only be valid
under the assumption that the methods of fear appeals and modelling do not interact in their
effectiveness. This is unlikely (in general; and in this case, specifically so, because modelling can also
be used to influence self-efficacy, a requirement for fear appeal effectiveness), and therefore, we will
need to include predictors to model these interactions. Although all types of each method can be
efficiently modelled with only 7 dummy variables, modelling the interaction between these types
requires an additional 96 dummy variables (one for each cell of a 12x8 matrix; see Table 3). Thus,
proper assessment of the relative effectiveness of these two methods of behaviour change requires a
meta-regression with 103 predictors (i.e. 7 dummy predictors plus 96 interaction predictors). Such an
analysis would only address two of the dozens of potential methods, only address two-way
interactions, and ignore the role of contextual factors such as population and behavioural domain.
Table 1: All possible combinations of satisfaction of the four parameters of effectiveness for fear
appeals.
Id Severity Susceptibility
Response efficacy
Self-
efficacy Illustrative description
1 0 0 0 0 No manipulation
2 0 0 0 1
3 0 0 1 0
4 0 0 1 1 Efficacy-only manipulation, e.g. guided practice; hardly a ‘fear’ appeal
5 0 1 0 0 Susceptibility-only manipulation, e.g. scenario-based risk information
6 0 1 0 1
7 0 1 1 0
8 0 1 1 1
9 1 0 0 0 Severity-only manipulation, e.g. most warning labels on tobacco packaging
10 1 0 0 1
11 1 0 1 0
12 1 0 1 1
13 1 1 0 0 Threat-only manipulation, e.g. a physician discusses test outcomes with a patient
14 1 1 0 1
15 1 1 1 0
Illustration of number of dummy variables required to model parameters of effectiveness
Supplemental material to “Everything should be as simple as possible, but no simpler: towards a protocol for accumulating evidence regarding the active content of health behaviour change interventions” by Peters, de Bruin & Crutzen in Health Psychology Review (2014)2 of 2
Table 2: All possible combinations of satisfaction of the four parameters of effectiveness for
modelling.
Id
Attention, remembrance
& skills Identification
Positive reinforcement
Coping
model Illustrative description
1 0 0 0 0 No manipulation 2 0 0 0 1 3 0 0 1 0 4 0 0 1 1 5 0 1 0 0 6 0 1 0 1 7 0 1 1 0 8 0 1 1 1
9 1 0 0 0 A soccer star effortlessly quits smoking without reinforcement
10 1 0 0 1
11 1 0 1 0 A famous singer explains that regular exercise is easy for her, and that it substantially improved her life
12 1 0 1 1
13 1 1 0 0 A peer educator explains how easy it was for him to negotiate condom use
14 1 1 0 1
15 1 1 1 0
16 1 1 1 1 Correctly applied modelling
Table 3: Dummy variables required to model the interaction between the 12 types of fear appeals and
the 8 types of modelling.
Modelling Id 1 9 10 11 12 13 14 15 16 Fe ar appeal s 1 0 0 0 0 0 0 0 0 0 5 0 1 1 1 1 1 1 1 1 6 0 1 1 1 1 1 1 1 1 7 0 1 1 1 1 1 1 1 1 8 0 1 1 1 1 1 1 1 1 9 0 1 1 1 1 1 1 1 1 10 0 1 1 1 1 1 1 1 1 11 0 1 1 1 1 1 1 1 1 12 0 1 1 1 1 1 1 1 1 13 0 1 1 1 1 1 1 1 1 14 0 1 1 1 1 1 1 1 1 15 0 1 1 1 1 1 1 1 1 16 0 1 1 1 1 1 1 1 1
Note: Id 1 means that the method is absent from an intervention; this situation is the reference situation, and thus no dummies are needed.